I guess it depends on whether you need the development environment or not. I did some voice-synthesis programming for DOS, Unix and Windows. The DOS and Unix versions were all in one codebase, totalling about 2000 lines of code including the assembly routines needed for DOS. It supported 3 driver variations under DOS, 5 under Linux and 2 under Solaris. The DOS executable would run on a 16MHz 286 with no skipping or drop-outs. The Linux executable would run on a 20MHz 386, and by the time you loaded it up enough to get noticeable skipping and drop-outs in the audio it was taking 10-20 seconds for console keystrokes to echo. The only development environment I needed was Emacs.
The Windows version used the DirectSound interface. I had to use the development environment because that was the only way to generate all the overhead code. Manually generated code totalled about 6000 lines to support just the Win32 and DirectX APIs (ie. no support for DOS or Win3.1 at all). Auto-generated code added several thousand lines to that. Performance... well, on a 200MHz Pentium Pro I could cause painful drop-outs in the audio just by moving the mouse around. As for development time, it took longer to develop the one Windows version than it took to develop the entire combined DOS/Unix one even with the development environment on Windows.
I think the main reason you don't see complex IDEs for Unix is that Unix programmers try to avoid the kinds of complex frameworks that make IDEs such a neccessity in the Windows world.
It will not stop your MP3s from playing. They'll just play in a protected address space. Or maybe they won't depending on your player software.
I think you misunderstand the use of the protected area. Your MP3s will be encrypted. the keys to decrypt them will be stored in the protected area. Only "trusted" programs will be allowed access to the protected area, so only "trusted" programs will be able to get the keys and decrypt the MP3s. One requirement for "trust" will be that the player provides no way to save the unencrypted datastream anywhere. Possibly it might not even send the stream to a sound card unless that sound card was also "trusted".
The fundamental problem isn't even the word "trust", it's who can trust the computer. This whole thing isn't intended to insure that you can trust your computer or the software on it. It's to insure that other people (eg. the RIAA, MPAA, Microsoft) can trust your computer. Trust it, that is, to do only what they tell it it can do and nothing else. If you wonder why MS would want that, think back a couple of years to their floating of the idea of annual subscriptions for Windows licenses. Now imagine the glee when they discover a way to guarantee that, if they impose that, you the user can't do a thing to bypass their check of whether you've paid or not because the hardware won't let you touch that data.
You're forgetting one thing, though: the Judge isn't leaving it up to the DoJ. Remember the one thing she wants before she'll approve the settlement: an explicit statement in the settlement that the court has the authority to issue orders regarding the implementation of the settlement. MS doesn't have to just deal with Bush and his lapdogs at the DoJ, they've got to deal with a judge who appears to be quite savvy about things.
No, as far as I can tell what she wants is either a) the ISP to start bouncing e-mail to closed accounts as soon as they're closed/suspended/whatever, or b) the ISP to allow access to the e-mail for as long as they choose to accept and not bounce it. What she objects to is the policy of accepting the mail as if everything's OK (thus not giving the sender any clue there's any problem) while simultaneously not allowing the recipient access to it (thus not giving the recipient any way of knowing who/what to respond to). To me her position seems emminently reasonable.
It's not quite so much of a catch as it first appears. They can't refuse to release it just because it's related to or involves security and such, they can only refuse if releasing it would compromise security. For instance, take domain authentication. If it's actually secure, knowing the communications protocol and encryption algorithms won't compromise it's security, only knowing the user's keys would. For MS to keep it undisclosed, they'd have to state basically that if you knew what went back and forth on the wire you could crack domain authentication without knowing anybody's keys. Ditto DRM, they can't refuse to release the info unless they openly claim that knowing how to let your files be protected but usable would break the entire system. They can be pushed pretty hard on these things in the press if they balk.
Thing is, KMart isn't selling the licenses. It's selling bluelight.com, the entity which licensed the software. It seems to me that the licenses should, under conventional law, remain with the entity that holds them when it's sold to someone else. For the licenses to stay with KMart, bluelight.com would have to transfer the licenses to KMart.
MS may have made a mistake pulling this during a bankruptcy proceeding where the judge has a lot of leeway in saying "This is the way it'll be.".
Except that the equivalent of an unlinked URL isn't an unpublished phone number, it's a published-but-unadvertised phone number: you don't go around telling people it's there, but Directory Assistance will cheerfully give it out if someone asks them because you haven't told them not to. Depending on people not calling DA and asking for your phone number seems, to me, to not show any intent to keep the information private.
I think your question can be answered by considering the difference between a URL and a link. A URL names a file on (typically) a Webserver. That URL exists as soon as the file exists. You can type it into the URL bar of a browser and access the file at that URL regardless of anything else.
A link, OTOH, is a tag in HTML that points to a URL and lets someone access that URL by clicking the link instead of having to type the URL in by hand. Links require URLs, but URLs don't require links.
Consider it like a phone number. Suppose a company gets a phone line and gets a regular number, not unlisted or unpublished or anything. Their phone number exists, right? Now, suppose they don't put that phone number in their advertising material or anything. They don't make it unlisted, they just don't mention it. Can they claim any sort of protection from people calling that number?
Only one problem with your GPL analysis. First off, if it's only their property, there's no GPL issue. If they wrote all the code, they can release it under any license they like. And if they included GPL'd code, it isn't their property. In that case, why should their desire to commercialize the code give them the right to ignore the license the owner of the GPL'd code put on it? In short, what gives them the right to use someone else's property any way they like regardless of the license terms on it?
You limit the potential infection zone to your directories, yes. As KLEZ proved, though, even a Unix user can do a lot of damage. File ownerships don't prevent a worm from sending out infected mail, or stealing your personal documents (the most incriminating/sensitive documents on a Unix system aren't likely to be owned by root), or even acting as a zombie for a remote-control DDoS (it can't forge IP header fields or use privileged ports, but it can still generate lots of traffic). We're not as wide-open as Windows, but many of the same dangers are there if we leave the door open.
Because opening a.exe with WINE means running it, which means there shouldn't be an entry in KMail's MIME-handler list telling it to open.exe files with WINE.
This isn't just limited to WINE, it can hit real Linux mail programs too if anyone ever writes a Linux/ELF virus attachment. Repeat after me, kids:
Executable MIME types have no place in a mail program!
None, never, no way. Mail program doesn't matter. OS doesn't matter. No mail program should ever, under any circumstances, execute anything attached to an e-mail message, period full stop. You should only execute things from people you trust, and one attribute of e-mail is that you don't even know if the From address is the real sender so how can you trust the message?
No, I'm not. Under the proposed rules, any Webcaster owes a minimum of $500/year to the RIAA regardless of whether they stream any songs on which they would owe royalties to the RIAA. Unless, maybe, they can account for every single second of every stream and what they played, and have signed waivers of CARP fees from every artist whose songs they've played, and even then it's not clear that they wouldn't owe the minimum.
Except that he still didn't buy PageRank from Google. Google assigned it to him based on relevance. If he suddenly starts linking to lots of non-relevant pages, why should he be suprised if Google downgrades him? He's no longer a source of high-relevancy links, after all.
Bob wants something, all right, but it's not for Google to be aware of their commodity. He wants Google to guarantee him a certain amount of that commodity that he can sell and profit from. Google declines to do that. Tough for Bob.
The only problem is that before Bob could have bought PageRank, Google would have had to have been selling it. They don't sell PageRank. They sell ad placements in other areas, but PageRank isn't something you can buy from them any more than you can buy a good review on a book from a reliable critic. All Google did was downgrade their estimate of the worth of references from him after they determined he was in the business of inflating the worth of pages he listed.
Yes, it's going to hurt his business. It's going to hurt the business of a book critic to have it advertised that you can buy a good review from him. The fault for the damage no more lies with Google than it does with the papers and shows that no longer rely on the critic's bought-and-paid-for reviews.
Actually it's in the bills themselves. If you read the text, there's no mention at all of an exemption from the royalties when you broadcast songs on which you hold the copyrights. The bills are written with the unstated assumption that the RIAA holds the copyrights on all songs and doesn't leave any outs for when they don't. So, any artist broadcasting their own music has to pay the royalties on every song, then go begging to the RIAA for a small slice of the money back. And if they aren't an RIAA member, they aren't entitled to the money.
Developers won't matter. When the legal department looks at the agreements and says "What are you doing letting someone else make a profit off our property without us getting a cut of it?", management will sit up and take notice.
We'll do what we've traditionally done: get paid to write software. I'd say about 80% of software is by it's nature not amenable to being widely distributed. For example, a point-of-sale system tied tightly into the pump-control, tank-monitoring and other hardware of a truckstop. Half a million or so lines of code, all told, and all of it so specific to one company's way of doing business that there's only a handful of other people who could use it without major modifications and customization. For all that, though, it's so critical to keeping the company running that abandoning it in favor of more generic solutions would be corporate suicide. It would simply cost too much in lost opportunities to have to wait 5 years for someone else to implement an idea, not to mention the costs of customizing it to match the way the company works (or alternatively changing the way the company works, but that's letting the tail wag the dog).
In that kind of situation, open-source is infrastructure. It's the generic code that handles the routine jobs and the well-known tasks so the programmers can concentrate on the critical parts that aren't generic.
There's one aspect of Open Source that Ballmer and his friends don't get yet. He talks about trying to adopt the open-source ideas to benefit Microsoft. That dooms him to failure right there. People don't contribute to open-source software to benefit someone else. They contribute to benefit themselves. They fix bugs and add features because they need that done. And the contribute it back because they've already benefited from previous contributions from other people. It's all aimed at the benefit of the customer/user. When anyone, whether they be Microsoft or Sun or whoever, sets up a similar system aimed to benefit someone other than the people actually doing the work, those people don't buy in and the whole thing kind of shambles off into oblivion.
If Ballmer wants to adopt open-source ideas, the first one is going to have to be "How can our users add to and change Windows to benefit themselves?". As long as "How can users add to Windows to benefit Microsoft?" takes priority, it'll fail.
Only problem with your outrage: the license specifically says that the part you bolded applies only to messages posted to message boards and bug reports posted to the bug-reporting tools. The first and last paragraphs explicitly leave source code, documentation and other material licensed only under the license the creator put on it and any license they grant to SourceForge.
One point you might make to the corporate types to get them to side with royalty-free is that it's in their interests. Their usual method of dealing with patents in their way is to find a patent they hold that the other guy infringes on and use that as leverage to get a no-cost cross-licensing agreement. A couple of big cases lately, eg. the JPEG stuff, have involved patents held by people whose sole product is the patents they hold. They don't make anything, therefore they don't make anything that could infringe on any other patents, so there's no reason for them to cross-license. More and more, the corporations are going to be dealing with patents held by people who the corporation won't have any leverage with. And as more patents are issued, a corporation will more and more often be on the wrong side of the equation, ie. they'll be the ones defending against patent enforcement instead of being the ones doing the enforcing. Royalty-free may cost them a bit on the patents they hold, but non-royalty-free would seem to potentially cost them a lot more on patents they don't hold and can't get a cross-licensing agreement on.
But the bill doesn't make EULAs illegal. It just makes ones that limit your rights to less than those you have under the law, without letting you negotiate them, illegal. To me this sounds perfectly sensible: the law sets the minimum rights, and nobody can be asked to give up their legal rights without an actual negotiation in which both sides have a say.
As for EULAs you have to agree to after purchase, those should be made illegal. Under the UCC contract's formed at purchase, and one side shouldn't be able to unilaterally change the terms of the contract after the fact.
Prosecuting under the law for this particular form of theft and trespass is exactly what the California AG is doing. They merely passed the law because, in the case of spam, it's impossible not to commit trespass and theft by doing it. Oh, some people might not mind, but you didn't know that when you did it.
As for the ISP, well, only if the ISP's TOS includes a "you must accept spam" clause. If it doesn't, then while the ISP might not have recourse the individual subscriber does because the spammer's still trespassing and taking up the subscriber's disk space and bandwidth without permission. If it were the ISP paying it might be different, but the ISP isn't paying and the subscriber has the bill from the ISP to prove it.
Slashdot Mirror
I guess it depends on whether you need the development environment or not. I did some voice-synthesis programming for DOS, Unix and Windows. The DOS and Unix versions were all in one codebase, totalling about 2000 lines of code including the assembly routines needed for DOS. It supported 3 driver variations under DOS, 5 under Linux and 2 under Solaris. The DOS executable would run on a 16MHz 286 with no skipping or drop-outs. The Linux executable would run on a 20MHz 386, and by the time you loaded it up enough to get noticeable skipping and drop-outs in the audio it was taking 10-20 seconds for console keystrokes to echo. The only development environment I needed was Emacs.
The Windows version used the DirectSound interface. I had to use the development environment because that was the only way to generate all the overhead code. Manually generated code totalled about 6000 lines to support just the Win32 and DirectX APIs (ie. no support for DOS or Win3.1 at all). Auto-generated code added several thousand lines to that. Performance... well, on a 200MHz Pentium Pro I could cause painful drop-outs in the audio just by moving the mouse around. As for development time, it took longer to develop the one Windows version than it took to develop the entire combined DOS/Unix one even with the development environment on Windows.
I think the main reason you don't see complex IDEs for Unix is that Unix programmers try to avoid the kinds of complex frameworks that make IDEs such a neccessity in the Windows world.
It will not stop your MP3s from playing. They'll just play in a protected address space. Or maybe they won't depending on your player software.
I think you misunderstand the use of the protected area. Your MP3s will be encrypted. the keys to decrypt them will be stored in the protected area. Only "trusted" programs will be allowed access to the protected area, so only "trusted" programs will be able to get the keys and decrypt the MP3s. One requirement for "trust" will be that the player provides no way to save the unencrypted datastream anywhere. Possibly it might not even send the stream to a sound card unless that sound card was also "trusted".
The fundamental problem isn't even the word "trust", it's who can trust the computer. This whole thing isn't intended to insure that you can trust your computer or the software on it. It's to insure that other people (eg. the RIAA, MPAA, Microsoft) can trust your computer. Trust it, that is, to do only what they tell it it can do and nothing else. If you wonder why MS would want that, think back a couple of years to their floating of the idea of annual subscriptions for Windows licenses. Now imagine the glee when they discover a way to guarantee that, if they impose that, you the user can't do a thing to bypass their check of whether you've paid or not because the hardware won't let you touch that data.
You're forgetting one thing, though: the Judge isn't leaving it up to the DoJ. Remember the one thing she wants before she'll approve the settlement: an explicit statement in the settlement that the court has the authority to issue orders regarding the implementation of the settlement. MS doesn't have to just deal with Bush and his lapdogs at the DoJ, they've got to deal with a judge who appears to be quite savvy about things.
No, as far as I can tell what she wants is either a) the ISP to start bouncing e-mail to closed accounts as soon as they're closed/suspended/whatever, or b) the ISP to allow access to the e-mail for as long as they choose to accept and not bounce it. What she objects to is the policy of accepting the mail as if everything's OK (thus not giving the sender any clue there's any problem) while simultaneously not allowing the recipient access to it (thus not giving the recipient any way of knowing who/what to respond to). To me her position seems emminently reasonable.
It's not quite so much of a catch as it first appears. They can't refuse to release it just because it's related to or involves security and such, they can only refuse if releasing it would compromise security. For instance, take domain authentication. If it's actually secure, knowing the communications protocol and encryption algorithms won't compromise it's security, only knowing the user's keys would. For MS to keep it undisclosed, they'd have to state basically that if you knew what went back and forth on the wire you could crack domain authentication without knowing anybody's keys. Ditto DRM, they can't refuse to release the info unless they openly claim that knowing how to let your files be protected but usable would break the entire system. They can be pushed pretty hard on these things in the press if they balk.
Thing is, KMart isn't selling the licenses. It's selling bluelight.com, the entity which licensed the software. It seems to me that the licenses should, under conventional law, remain with the entity that holds them when it's sold to someone else. For the licenses to stay with KMart, bluelight.com would have to transfer the licenses to KMart.
MS may have made a mistake pulling this during a bankruptcy proceeding where the judge has a lot of leeway in saying "This is the way it'll be.".
Except that the equivalent of an unlinked URL isn't an unpublished phone number, it's a published-but-unadvertised phone number: you don't go around telling people it's there, but Directory Assistance will cheerfully give it out if someone asks them because you haven't told them not to. Depending on people not calling DA and asking for your phone number seems, to me, to not show any intent to keep the information private.
I think your question can be answered by considering the difference between a URL and a link. A URL names a file on (typically) a Webserver. That URL exists as soon as the file exists. You can type it into the URL bar of a browser and access the file at that URL regardless of anything else.
A link, OTOH, is a tag in HTML that points to a URL and lets someone access that URL by clicking the link instead of having to type the URL in by hand. Links require URLs, but URLs don't require links.
Consider it like a phone number. Suppose a company gets a phone line and gets a regular number, not unlisted or unpublished or anything. Their phone number exists, right? Now, suppose they don't put that phone number in their advertising material or anything. They don't make it unlisted, they just don't mention it. Can they claim any sort of protection from people calling that number?
Only one problem with your GPL analysis. First off, if it's only their property, there's no GPL issue. If they wrote all the code, they can release it under any license they like. And if they included GPL'd code, it isn't their property. In that case, why should their desire to commercialize the code give them the right to ignore the license the owner of the GPL'd code put on it? In short, what gives them the right to use someone else's property any way they like regardless of the license terms on it?
You limit the potential infection zone to your directories, yes. As KLEZ proved, though, even a Unix user can do a lot of damage. File ownerships don't prevent a worm from sending out infected mail, or stealing your personal documents (the most incriminating/sensitive documents on a Unix system aren't likely to be owned by root), or even acting as a zombie for a remote-control DDoS (it can't forge IP header fields or use privileged ports, but it can still generate lots of traffic). We're not as wide-open as Windows, but many of the same dangers are there if we leave the door open.
Because opening a .exe with WINE means running it, which means there shouldn't be an entry in KMail's MIME-handler list telling it to open .exe files with WINE.
This isn't just limited to WINE, it can hit real Linux mail programs too if anyone ever writes a Linux/ELF virus attachment. Repeat after me, kids:
Executable MIME types have no place in a mail program!
None, never, no way. Mail program doesn't matter. OS doesn't matter. No mail program should ever, under any circumstances, execute anything attached to an e-mail message, period full stop. You should only execute things from people you trust, and one attribute of e-mail is that you don't even know if the From address is the real sender so how can you trust the message?
No, I'm not. Under the proposed rules, any Webcaster owes a minimum of $500/year to the RIAA regardless of whether they stream any songs on which they would owe royalties to the RIAA. Unless, maybe, they can account for every single second of every stream and what they played, and have signed waivers of CARP fees from every artist whose songs they've played, and even then it's not clear that they wouldn't owe the minimum.
Except that he still didn't buy PageRank from Google. Google assigned it to him based on relevance. If he suddenly starts linking to lots of non-relevant pages, why should he be suprised if Google downgrades him? He's no longer a source of high-relevancy links, after all.
Bob wants something, all right, but it's not for Google to be aware of their commodity. He wants Google to guarantee him a certain amount of that commodity that he can sell and profit from. Google declines to do that. Tough for Bob.
The only problem is that before Bob could have bought PageRank, Google would have had to have been selling it. They don't sell PageRank. They sell ad placements in other areas, but PageRank isn't something you can buy from them any more than you can buy a good review on a book from a reliable critic. All Google did was downgrade their estimate of the worth of references from him after they determined he was in the business of inflating the worth of pages he listed.
Yes, it's going to hurt his business. It's going to hurt the business of a book critic to have it advertised that you can buy a good review from him. The fault for the damage no more lies with Google than it does with the papers and shows that no longer rely on the critic's bought-and-paid-for reviews.
Actually it's in the bills themselves. If you read the text, there's no mention at all of an exemption from the royalties when you broadcast songs on which you hold the copyrights. The bills are written with the unstated assumption that the RIAA holds the copyrights on all songs and doesn't leave any outs for when they don't. So, any artist broadcasting their own music has to pay the royalties on every song, then go begging to the RIAA for a small slice of the money back. And if they aren't an RIAA member, they aren't entitled to the money.
Developers won't matter. When the legal department looks at the agreements and says "What are you doing letting someone else make a profit off our property without us getting a cut of it?", management will sit up and take notice.
Oh, I get it, all right. And you've just repeated what I said about why MS's attempts to do this won't work.
We'll do what we've traditionally done: get paid to write software. I'd say about 80% of software is by it's nature not amenable to being widely distributed. For example, a point-of-sale system tied tightly into the pump-control, tank-monitoring and other hardware of a truckstop. Half a million or so lines of code, all told, and all of it so specific to one company's way of doing business that there's only a handful of other people who could use it without major modifications and customization. For all that, though, it's so critical to keeping the company running that abandoning it in favor of more generic solutions would be corporate suicide. It would simply cost too much in lost opportunities to have to wait 5 years for someone else to implement an idea, not to mention the costs of customizing it to match the way the company works (or alternatively changing the way the company works, but that's letting the tail wag the dog).
In that kind of situation, open-source is infrastructure. It's the generic code that handles the routine jobs and the well-known tasks so the programmers can concentrate on the critical parts that aren't generic.
There's one aspect of Open Source that Ballmer and his friends don't get yet. He talks about trying to adopt the open-source ideas to benefit Microsoft. That dooms him to failure right there. People don't contribute to open-source software to benefit someone else. They contribute to benefit themselves. They fix bugs and add features because they need that done. And the contribute it back because they've already benefited from previous contributions from other people. It's all aimed at the benefit of the customer/user. When anyone, whether they be Microsoft or Sun or whoever, sets up a similar system aimed to benefit someone other than the people actually doing the work, those people don't buy in and the whole thing kind of shambles off into oblivion.
If Ballmer wants to adopt open-source ideas, the first one is going to have to be "How can our users add to and change Windows to benefit themselves?". As long as "How can users add to Windows to benefit Microsoft?" takes priority, it'll fail.
Not neccesarily. But then, if they were so reliable they wouldn't be getting returned under warranty because they broke, now would they? :)
Only problem with your outrage: the license specifically says that the part you bolded applies only to messages posted to message boards and bug reports posted to the bug-reporting tools. The first and last paragraphs explicitly leave source code, documentation and other material licensed only under the license the creator put on it and any license they grant to SourceForge.
Thanks for the representation, Bruce.
One point you might make to the corporate types to get them to side with royalty-free is that it's in their interests. Their usual method of dealing with patents in their way is to find a patent they hold that the other guy infringes on and use that as leverage to get a no-cost cross-licensing agreement. A couple of big cases lately, eg. the JPEG stuff, have involved patents held by people whose sole product is the patents they hold. They don't make anything, therefore they don't make anything that could infringe on any other patents, so there's no reason for them to cross-license. More and more, the corporations are going to be dealing with patents held by people who the corporation won't have any leverage with. And as more patents are issued, a corporation will more and more often be on the wrong side of the equation, ie. they'll be the ones defending against patent enforcement instead of being the ones doing the enforcing. Royalty-free may cost them a bit on the patents they hold, but non-royalty-free would seem to potentially cost them a lot more on patents they don't hold and can't get a cross-licensing agreement on.
But the bill doesn't make EULAs illegal. It just makes ones that limit your rights to less than those you have under the law, without letting you negotiate them, illegal. To me this sounds perfectly sensible: the law sets the minimum rights, and nobody can be asked to give up their legal rights without an actual negotiation in which both sides have a say.
As for EULAs you have to agree to after purchase, those should be made illegal. Under the UCC contract's formed at purchase, and one side shouldn't be able to unilaterally change the terms of the contract after the fact.
Prosecuting under the law for this particular form of theft and trespass is exactly what the California AG is doing. They merely passed the law because, in the case of spam, it's impossible not to commit trespass and theft by doing it. Oh, some people might not mind, but you didn't know that when you did it.
As for the ISP, well, only if the ISP's TOS includes a "you must accept spam" clause. If it doesn't, then while the ISP might not have recourse the individual subscriber does because the spammer's still trespassing and taking up the subscriber's disk space and bandwidth without permission. If it were the ISP paying it might be different, but the ISP isn't paying and the subscriber has the bill from the ISP to prove it.